This recently happened to me...

I would first reboot from the Live CD, and run full checks on the partitions:

You can do this from the GUI, but if you do it from the console (as root), the command is:

# fsck -cf /dev/hda1

Replace hda1 by your actual partition name, and repeat for each partition.

...reboot your system and see if it goes back into the GUI. You did get the GUI to load properly during your install right?

If you still don't get the GUI to come up, let us know what messages you have on your screen. If you have the "login" screen with a message above it stating "Press ctrl-alt-f7 (or f8...)" (don't remember exact verbiage), then try pressing ctrl-alt-f7 together to take you to the GUI login screen. Did you get there?

If that doesn't work, you may need to reload X Window. But I'm not sure why you would have to do this. If necessary, use the Mepis Live CD, choose to Install, and then choose the Reinstall X Config option.

And if your son wants to "help" you again, tell him to get his own computer!

Jon

The next time you install Linux on that, or any other box, be sure to select the ReiserFS. Using ReiserFS is almost as good as having an UPS connected to your box.

I've lost count of the number of times a sudden power outage caused my box to reboot without a normal shutdown. I also use ReiserFS on my Linux boxes at work. Since I first started using ReiserFS on SUSE 6.4, about 6 years ago, on maybe a dozen different boxes, I have NEVER lost data or had the journaling system fail to recover after an abend.

--
GreyGeek

Ditto that, GG.

We had a breaker trip the other night and, when my wife's computer, and mine, were powered-back-up (simultaneously) I was back at my work within 2 or 3 minutes; while she had to wait about 30 minutes to just be able to log back into her windoze machine.

Reiser is pretty darn good.

"You done filled my wagon chock-full to burstin' with nonessentials. Look at this: Cinnamon, oregano, ci-lantro...What in the cock-a-doodle is ci-lantro?"
"I got your four basic food groups! Beans, bacon, whiskey and lard!"

Hmm.... Food for thought. I've already installed my OS though. I wonder if I can convert to ReiserFS from EXT3 like you can from FAT32 to NTFS.

"Windows [n.]
A thirty-two bit extension and GUI shell to a sixteen bit patch to an eight bit operating system originally coded for a four bit microprocessor and sold by a two-bit company that can't stand one bit of competition."
(Anonymous USEnet post

you can't just convert from one to the other
my last experiences with Reiserfs were not good, with certain combinations of activity causing kernel panic on a regular basis on a SuSE 7.x system.
my dying machine has crashed too many times and will no longer let me in with ssh - maybe if i too had reiser not ext3 on it i would have survived the last crashes.

The fact that you can convert ext2 to ext3 WITHOUT reformatting your HD tells you something about ext3 and its ability to journal data: ext2 is still the underlying fs.

I tried using ext3 a year ago just to compare it with ReiserFS. I installed it on a box that had been running MEPIS with ReiserFS for about a year without problems. On the FIRST reboot after the install fsck interrupted the install and said the HD had errors. I let it run and 30 MINUTES later the boot continued without errors. That was only a 30GB HD. I shudder at the thought of running fsck on a 260GB HD!

I was curious about why ext3 should require fsck if it using a journaling system. After consulting the code I learned that ext3 is just "tacked onto" ext2 and has a counter that forces an fsck every 100 reboots (default) or at what ever number you set when you run the formatter.

I immediately reinstalled MEPIS but used ReiserFS. I had no further problems with that box.

--
GreyGeek

I thought others might be interested in reading about ReiserFS. I'd never heard of it before. Interesting.

In Depth Look into ReiserFS

With all due respect, I ran reiserfs on my laptops and my backup systems for many years. I first heard of it and began using it when I used SuSE as my OS. And for the last few years, when I heard that Reiser4 was coming out, I have been excited to see what would happen.

And I shall not bring up the unfortunate situation with Hans Reiser

However, there is one very bad trait with reiserfs. I have had to perform massive restores from one computer to another (due to my constant mucking about and adding/deleting/formatting... to learn stuff). Because reiserfs builds the journal and underlying file structure "as needed", rather than the pre-configuring of inodes and such used by ext2/3, fast, large restores can mess up big time.

If I copied all the data from remote:/home to local:/home for instance and local:/home was on a reiserfs partition, because so much data was being streamed over the network and being written so fast, reiserfs would choke. It would start throwing error stupid error messages along the lines of "such-and-such file cannot be created because it doesn't already exist...". It was as if the internal indexing was being created before the file actually got written, and then caused a conflict. To its credit, however, by performing multiple "rsyncs" and deletions of the "offending non-existent files" I was eventually able to get everything restored.

However, with my new system, I have put /home on a jfs file system. The "root" partition has been set up as ext3.

The reason ext3 is a journal on top of ext2 is for backward compatibility. A rather good, conservative way of approaching a production problem. If the journal blows up, just delete it, and rebuild it.

And saintj0n, it is possible to convert one file system to another but it is dangerous. There is a package available in the repositories called convertfs (apt-get install convertfs). I have used it to convert to/from reiserfs, ext2/3, and jfs. But I've had only a 50% success rate. When it converts, everything is wonderful. When it doesn't I have to reformat

I do not recommend convertfs unless you have a backup of all your good stuff!

Oh Cramp, thanks for the informative link describing reiserfs!

Jon

To Liz's experiences with ext3:

SuSE: is something I tried a while back ('round 2004, I believe), but it didn't strike me as a good match for my situation. I have left it alone since, so commenting as to your experiences would be foolish on my part.

ext2/3 & trashed crash recovery: That's exactly what drove me to go with Reiser, although I do realize that there are certain circumstances where Reiser will act-up.

Hans Reiser: No comment. Will keep an eye on developments.

Jon: I've never tried performing massive restores in the way that you've outlined. I've usually dd-ed from one device to the other on the local host, but I'm one to pull HDs and move them around liberally, so I may naturally be dancing around the conditions to create the problem.

Am I reading you correctly as saying the network bottleneck was a contributing factor in what you've seen?

Something to investigate. I'm currently giving XFS the eye, but don't have any issues pushing me away from ReiserFS.

"You done filled my wagon chock-full to burstin' with nonessentials. Look at this: Cinnamon, oregano, ci-lantro...What in the cock-a-doodle is ci-lantro?"
"I got your four basic food groups! Beans, bacon, whiskey and lard!"

The nice new box I set up with reiserfs but it was a no-go. Perhaps the problem was similar to Jon's rapid restore problem.
So I reformatted as ext3 and am running fine.

When I installed MEPIS last night, I tried to select ReiserFS, and it installed, but when it came "GRUB" time in the install, it refused to install it into the MEPIS partition. Why is that?

I restarted installation, and I'm working with "ext3" for the moment, but what should I do to remedy that?

ext3 will be fine.
the object is to get the system running, and you can expect any of the linux journalled filesystems to work well.

I think this is best blamed on the Coriolis Effect.

My occupation?
Well, computer geek-stuff, mostly. I could tell you all about it; but, then I would have to delete you.

By gum EnigmaOne I think you're on to something! If the hard drive spins clockwise and is used in the southern hemisphere where the Coriolis effect causes water to go down the drain in a counter-clockwise direction, it could cause the magnetic bits on the platter to want to reverse their charge and fly off the platter, sticking to the inside of the hard drive case. I think the only solution is to put the hard drive into the case upside down.

Jon

Interesting discussion.

I've actually been thinking about switching to ext3 (I'm currently using ReiserFS).

Why?

I don't know much about Linux file systems. But, I remembered reading an article recently about Suse (before the Novell/MS fiasco) deciding to make ext3 the default, because ReiserFS was considered to be more of a "dead end" (my interpretation) file system and more developers would be working on ext3 in the future.

Here is one of the articles about it:

http://www.linux.com/article.pl?sid=06/10/12/2120204

Make sure to read the note to the Suse mailing list about it, since it goes into more detail:

http://lists.opensuse.org/opensuse-factory/2006-09/msg00542.html

I also noticed that some of the paritition managers don't let you resize ReiserFS.

Then, more recently when someone had a partition problem, I noticed that utilities like Photorec have some issues with ReiserFS. Here's a quote:

http://www.cgsecurity.org/wiki/PhotoRec#Filesystems

So, I've been thinking that the "safer bet" may be ext3.

If I had not noticed this thread, I may have even switched to it as soon as this coming week. Now, I guess I'll need to do a bit more research.

Jim C.

Hi Jim,

Hope you enjoy Serendipity.

To the discussion:
Back "in the day" (don't ask me why we start to say that when we hit our mid forties...we just do. Like Ken Starks says, it's like looking over our glasses.) I used to buy SnapOn tools; and was occasionally asked why I spent the greater money for a tool that had no better warranty than Sears Craftsman hand tools. My response was always along the lines of preferring the warranty that I'd never need, rather than the warranty I would end up using.

I have yet to have a problem with ReiserFS, but have had some lost data with both ext2 and ext3. I can't afford that, so I'll stick with Reiser until it lets me down.

That might be of help to you, or it may add more mud to the water. If the latter is true, I'll fall back on "It's the thought that counts."

My occupation?
Well, computer geek-stuff, mostly. I could tell you all about it; but, then I would have to delete you.

I honestly didn't realize that I had asked about it on your site. Or, did you just notice my post?

I've got my wife's laptop setup with ext3 (and it's been running that way for a long time now without any issues). Her battery died a long time ago. lol So, when the power blinks, her laptop reboots (not attached to an UPS). It's recovered every time we've had a power failure without issue.

But, there's probably nothing much going on that would cause corruption. It's not like she's doing a lot of writing to disk.

My PC is using ReiserFS. I don't have anything going on with it much that would cause much write activity either (yet). So, I don't know how reliable it is.

I guess I should just do some tests. I've got a few PCs sitting around that I've been meaning to play with.

Most of my experience with file systems comes from Netware (mostly 3.x). I've been known to demonstrate how great Novell's TTS (Transaction Tracking System) worked by pounding on large (for their time) shared database files containing millions of records, using multiple workstations writing to them (causing the server CPU utilization to be pretty much maxed out). Then, I'd deliberately yank the power and recover from it, and I never needed to use vrepair while demonstrating it (knock on wood) lol

As long as the files were flagged as Transactional, you'd never lose one, as the OS would back out the incomplete transactions when remounting the volume. Most people didn't even realize that you could flag files as transactional with Netware (and the performance hit was really pretty negligible). It had some downsides (for example, you couldn't delete a file flagged that way without removing the flag). But, it sure made your databases more reliable in the event of crashes.

I never lost one, never needing to rebuild any indexes (which would take a very long time for the file sizes at that time), or needed to restore anything from backup, running some mission critical systems 24x7 for a major long distance company some years back over about a 6 year period.

Ditto while running bunches of Netware based servers for a Tranportaton Related Company. We mirrored the drives (external SCSIs), and if anything other than a drive failed, we had identical servers sitting next to each other.

Move the drive cables to the backup box, fire it up, and you're back in business as soon as the server booted and volumes remounted. I've also had my share of experience with SFT III (but, the performance hit made that impractical for some of the stuff we did).

So, I should put together some similar tests and play with the file system types available for Linux to see how well the journaling actually works before deciding on what's best for long term use.

It gives me something to do.